﻿* Encoding: UTF-8.
*Logistic regression models*
*Pandemic effect on screening*

*Backwards selection--always keeping in region, practice type, and provider type, but iteratively dropping anything else with a p value exceeding .10

LOGISTIC REGRESSION VARIABLES PandemEffectonScreening01
  /METHOD=ENTER Age_4cat Race1 Ethnicity Gender_TransOtherasFemale Region1 ProviderType1 
    PracticeType 
  /CONTRAST (Age_4cat)=Indicator
  /CONTRAST (Race1)=Indicator
  /CONTRAST (Ethnicity)=Indicator
  /CONTRAST (Gender_TransOtherasFemale)=Indicator
  /CONTRAST (Region1)=Indicator
  /CONTRAST (ProviderType1)=Indicator
  /CONTRAST (PracticeType)=Indicator
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

*In above model, gender has highest p value so removing from below.

LOGISTIC REGRESSION VARIABLES PandemEffectonScreening01
  /METHOD=ENTER Age_4cat Race1 Ethnicity Region1 ProviderType1 
    PracticeType 
  /CONTRAST (Age_4cat)=Indicator
  /CONTRAST (Race1)=Indicator
  /CONTRAST (Ethnicity)=Indicator
  /CONTRAST (Region1)=Indicator
  /CONTRAST (ProviderType1)=Indicator
  /CONTRAST (PracticeType)=Indicator
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

*next removing ethnicity* This was the final model for screening. 

LOGISTIC REGRESSION VARIABLES PandemEffectonScreening01
  /METHOD=ENTER Age_4cat Race1 Region1 ProviderType1 
    PracticeType 
  /CONTRAST (Age_4cat)=Indicator
  /CONTRAST (Race1)=Indicator
  /CONTRAST (Region1)=Indicator
  /CONTRAST (ProviderType1)=Indicator
  /CONTRAST (PracticeType)=Indicator
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).


*Now colpo models* 

LOGISTIC REGRESSION VARIABLES PandemicEffectonColpo01
  /METHOD=ENTER Age_4cat Race1 Ethnicity Gender_TransOtherasFemale Region1 ProviderType1 
    PracticeType 
  /CONTRAST (Age_4cat)=Indicator
  /CONTRAST (Race1)=Indicator
  /CONTRAST (Ethnicity)=Indicator
  /CONTRAST (Gender_TransOtherasFemale)=Indicator
  /CONTRAST (Region1)=Indicator
  /CONTRAST (ProviderType1)=Indicator
  /CONTRAST (PracticeType)=Indicator
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

*Removing ethnicity from next model*

LOGISTIC REGRESSION VARIABLES PandemicEffectonColpo01
  /METHOD=ENTER Age_4cat Race1 Gender_TransOtherasFemale Region1 ProviderType1 
    PracticeType 
  /CONTRAST (Age_4cat)=Indicator
  /CONTRAST (Race1)=Indicator
  /CONTRAST (Gender_TransOtherasFemale)=Indicator
  /CONTRAST (Region1)=Indicator
  /CONTRAST (ProviderType1)=Indicator
  /CONTRAST (PracticeType)=Indicator
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).


*Removing race from next model* 

LOGISTIC REGRESSION VARIABLES PandemicEffectonColpo01
  /METHOD=ENTER Age_4cat Gender_TransOtherasFemale Region1 ProviderType1 
    PracticeType 
  /CONTRAST (Age_4cat)=Indicator
  /CONTRAST (Gender_TransOtherasFemale)=Indicator
  /CONTRAST (Region1)=Indicator
  /CONTRAST (ProviderType1)=Indicator
  /CONTRAST (PracticeType)=Indicator
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

*Removing age from next model* 

LOGISTIC REGRESSION VARIABLES PandemicEffectonColpo01
  /METHOD=ENTER Gender_TransOtherasFemale Region1 ProviderType1 
    PracticeType 
  /CONTRAST (Gender_TransOtherasFemale)=Indicator
  /CONTRAST (Region1)=Indicator
  /CONTRAST (ProviderType1)=Indicator
  /CONTRAST (PracticeType)=Indicator
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).


*Crosstabs* 

*Screening*

CROSSTABS
  /TABLES=Age_4cat Race1 Ethnicity Gender_TransOtherasFemale Region1 ProviderType1 PracticeType BY 
    PandemEffectonScreening01
  /FORMAT=AVALUE TABLES
  /STATISTICS=CHISQ 
  /CELLS=COUNT ROW COLUMN TOTAL 
  /COUNT ROUND CELL.


*Colpo* 

CROSSTABS
  /TABLES=Age_4cat Race1 Ethnicity Gender_TransOtherasFemale Region1 ProviderType1 PracticeType BY 
    PandemicEffectonColpo01
  /FORMAT=AVALUE TABLES
  /STATISTICS=CHISQ 
  /CELLS=COUNT ROW COLUMN TOTAL 
  /COUNT ROUND CELL.
